Millennium Bug and the Internet: analysis of possible problems.

Giovanni Prignano, President of Assonet – Confesercenti, President of Flashnet Spa and member of the Comitato Anno 2000

I have been invited to illustrate the implications for the Internet of the Year 2000 Issue: in this brief analysis, I shall attempt to convey my point of view as clearly as I possibly can.

In the current Y2K impact analytic scenario, much apprehension is generated by the alleged failure, more or less widspread, of the World Wide Web at the fatidical millennium roll over.

It is important, to better proceed with my analysis, to grasp how the "Net" uses its information resources to operate the service.

Let us first point out that all of the equipment used within the Net are of a high technological content.

Therefore, the continuous, almost daily technological advancements, are per se a guarantee that the most recent standards, including new date change specifications, are continually updated.

Among the countless problems that an operatore is to face, it is indeed the constant, inevitable necessity of equipment upgrading to enhance market competitiveness.

Moreover, the Internet represents an ideal ground wherein to confront and analyze the year 2000 Issue: here, a reasonable optimism obviously arises, mostly maintained by experiences as reported and compared within the numerous discussion groups which have been opened for this purpose.
(alt.talk.year2000; comp.software.year-2000)

Let us now review the technological aspect of the issue:

First of all, let us distinguish the main components of the Internet's neuronic structure.

There are, obviously, other components, but in our case we shall only consider those that we purposedly define as "primary".

All information that flows through the Net is codified into "packets", according to a specific protocol, the TCP-IP.

These packets circulate by means of various equipment which, in turn, contain instructions, in the form of software.

We then find systems for packet routing, others for data storage, others for data processing, others still for service access authorization (authentication).

Modern shelf information technology, as already underscored, is by now used by most international providers. Very frequently, they come as autonomous equipment which do not depend on associated computer systems.

As, for example, the Routers, systems that handle the routing of IP packets, other than physically linking digital connections between various exchanges.

These systems, which embed powerful microprocessors, are equipped with a software located into non-volatile memory (firmware) which, in most cases, can be upgraded online from remote location.

For some years, manufacturers have been providing customers with the necessary upgrades to avoid that, in the very few instances wherein the software handles date field, processing problems may surface and consequently affect correct routing.

The same logic could be applied to user's modem call clearing systems. System operation is based largely upon a passive portion (access server) which handles only the physical user's modem connection, and an intelligent portion (authentication server) which clears user's access to the Net.

These authentication servers, a real network nerve centre, are in absolute amongst the most controlled and managed network components.

If it is true that their operation is based upon the use of software applications that are potentially exposed to Y2K, it is otherwise correct to assume that system managements have already adopted the necessary problem solving measures since, as of beginning of this year, software application had to handle the date field in the new format, having the expiry date of annual access contracts gone past December, 1999.

The risk of having to explain to million of users why they are denied access from January 1, 2000, may indeed prove a very convincing deterrent for providers.

Systems which handle data storage, are evidently classifiable as processors, and as such exposed to problems relating to correct date field interpretation. The problem is therefore mandated to the software being used, but even in this case, the same logic may be applied to the rest of the technological structure: new version of the software, are indeed released with almost daily frequence.

We then have other processors which, being designed for more complex use, are equipped with most various software: in such cases we may hold doubts upon certain service correct operation, likely non-mission-critical, albeit however important.

We know well, as users, how frustrating it may become not being able to access our mailbox or any other web service. Besides it being a serious risk for the many businesses that by now use these services.

As far as electronic mail is concerned, we recall that an international agreement exists which all networks must refer to, and adhere to the rules issued therefrom. (it may be found as RFC at: http://sunsite.hr/rfc/index_fr.html )

Under this international agreement, mail technical management and coordination is carried out by detailing each time a uniform solution to the various problems.

To this regard, suggested corrections to e-mail options in view of the oncoming year end, have been around for some time.

A different analysis is required for e-commerce, a sector in which our Country, though heavily behind, is beginning to show a significant interest.

Online sale service is based upon a very close interaction among several entities: the Customer whom places the order, the Banking service that receives and aknowledges the payment, the supplier whom begins delivery procedures upon credit receipt.

Current systems, which mostly use credit cards as a payment medium, are obviously the most vulnerable showing possible criticality in such date management instances as credit card expiry date.

In fact, date is used everywhere: order date, shipping date, delivery date, etc.

Once again, without being excessively optimistic, we may rest relatively assured.

Indeed, since last January, credit cards and transactions with date field beyond December, 1999, have been in circulation without affecting normal current operations; an evidence of experimented problem analysis.

To conclude, we can state that the Internet, even thanks to effort and corrections, is way far from the catastrophic pictures which it has been framed in.

Web development, as service complexity and extent require, has always been the domain of highly qualified technical staff with an almost obsessive attention for the tinyest bug.

Now that the Internet has also become a business for large corporate, whether simple users or Operators, the level of attention is nearing its climax.

From what I have said, even in its necessary simplicity, I feel I can reassure that the possibilities of an unlikely failure of a modern Communication Highway, as the Internet is, have been reduced to the minimum.

Thanks for your attention.